Yelp’s Chad Richard On The Current State — And Near Future — Of Voice Activation

For the past several years, local digital guide Yelp has been working to move beyond being perceived as a “reviews site” to a platform that help make transactions between consumers and businesses.

For example, its Request-A-Quote feature, which lets Yelp users get the price of services before making a purchase via the guide platform, saw volume rise almost 30 percent over the past year.

Before that, Yelp expanded its restaurant services beyond its SeatMe and Yelp Now reservations tools with the purchase of former partner Nowait, a mobile platform that lets consumers virtually hold their place in line at casual dining establishments.

Yelp’s transaction business also was rounded out with last year’s $20 million purchase of location-based loyalty and retargeting platform, Turnstyle, which runs an in-store platform that then connects marketing services to consumers’ phones at 3,500 business places.

In a conversation at November’s Yext’s Onward 2017 conference with Yext President and Chief Revenue Officer Jim Steele,  Yelp COO Jed Nachman discussed  how the role of Connected Intelligence systems that power voice activated assistants and chatbots dovetails perfectly with the trajectory the 14-year-old company has taken.

“For voice and chat, you have to have the data to handle real-world interaction,” Nachman said of the company’s Yelp Knowledge, a tool that analyzes businesses’ reviews to help  understand the experience at specific locations.

To elaborate on how Yelp sees the rise of voice activation, we caught up with Chad Richard, Yelp’s SVP, Business and Corporate Development. Richard joined Yelp in 2015.

Before that, he spent six years at Apple as senior director of Worldwide Product Marketing focused on the Cupertino company’s operating systems and internet services, which included acquiring and building up the first mass market voice-enabled assistant, Siri.

GeoMarketing: What’s the state of voice activation and what does it mean for Yelp?

Chad Richard: Voice activation is a rapidly evolving paradigm of human-computer interaction that I’m pretty excited about. The utility of voice assistants and voice activation is especially high when people are on the go, and high quality local content is obviously a vital component to those mobile moments, so Yelp has a big role and opportunity here.

How does the Yelp Fusion API program reflect the influence of voice activation?

Yelp Fusion is a set of APIs and customized feeds for partners and independent developers who want to integrate Yelp content into their apps, websites, and services. Voice-activated devices and virtual assistants is a very popular Yelp Fusion use case and we’re already seeing some cool applications with our data.

How does Yelp work with other voice-activated platforms?

Yelp has partnered with Apple and been integrated into Siri since its launch in 2011. We also supply local data and content to Amazon for Alexa. So, products like Echo and Fire TV products are enabled with Yelp content. We work with Microsoft on Cortana and also collaborate with companies like Hound, both in their Hound products as well as Houndify, their platform where developers can develop their own voice activated applications that are Yelp enabled.

We work with Nuance, which has been part of the roots of all this, from the actual NLP standpoint to understand what people are actually saying. We’re also working with Samsung Voice, in coordination with Viv, a startup launched by the creators of Siri after they left Apple. Samsung bought Viv and it now powers Samsung’s personal assistant, “Bixby.”

So, even at this early stage, Yelp is virtually omnipresent in all the voice ecosystems. We’re powering voice experiences that are really great when you’re on-the-go, making queries to your smartphone, and even on your Echo when you are hanging out at home. You’re already using Yelp when talking to Bixby or Siri on your Android or iPhone.

The other place we’re seeing voice activation starting to grow is in the car. And that’s not just through Apple’s CarPlay, where you’re using Siri. It’s also native in dashboard experiences. So we’re already working with a lot of auto manufacturers via the Yelp Fusion API, and as those interfaces evolve to offer voice, Yelp will help to deliver our great local content as well.

Yelp’s always been pretty successful at getting our content integrated into car dashes, from BMW, Mercedes, Volvo, Lexus, Toyota, Honda, just to name a few automakers. Most of the auto manufacturers are working with us one way or another, either directly or through integrates like Harmon, or others selling into the auto space. In addition to finding great restaurants on-the-go, drivers can also reserve a table via Yelp Reservations so that there’s a table waiting for them when they arrive. Lastly, we’ve made it possible for drivers to “get in line” remotely at a restaurant through our acquisition of Nowait, which is now a part of Yelp Reservations.

How does Yelp serve as the bridge between brands and voice-activated devices?

When you are using voice enabled experience you want answers as opposed to a traditional search experience where you have a high tolerance for scanning various search results to find what you were really looking for. Yelp is a trusted source for high intent searches because we are able to deliver accurate information that’s highly relevant. This serves brands and businesses well by connecting them directly with the consumers via voice services.

How do you perceive the challenges for brands when it comes to using voice-activation?

One of the tough things about voice activation is for the assistant to really understand what consumers are really asking for. It means having as much content about the user as possible — such as identity, location and preferences – to having speech pattern technologies that can break down the query into specific nouns. If someone says  “Hong Kong café,” are they talking about a café in Hong Kong? Or are they talking about the Hong Kong Café down the street?

There are a million examples of how disambiguation becomes extremely important for these assistants to be smart and efficient for users.

There is the natural language processing and identity extraction side of all of this but its further complicated by the fact that these assistants are built with a broad range of data sources and therefore don’t just rely on Yelp data. We work closely with partners to help ensure they know when a question is best answered by Yelp.

How do you expect voice activation to shape marketing at the local level?

For us, voice represents an incredibly exciting human-to-computer interaction capability where local data is highly relevant. If you think about what Yelp’s been focused on for 14 years now — we’ve been connecting people with great local businesses.

But over the last few years, we have been really focused on getting transactional with it. And what’s cool about voice and Yelp’s role in this trend is the evolution from helping you find great restaurants and places to get a haircut, to actually being able to book a table or order pickup or delivery at that restaurant. Or go ahead and schedule that appointment at the hair salon. Yelp not only powers the discovery process, but it can power the purchase actions that follow.

From a marketing perspective, is the use of voice as opposed to text that different?

It’s not just about discovery with voice. Whereas text is offering a series of options for clicking, voice is about driving transactions. The consumer has this very clear intent that we can actually activate.

We’re seeing two types of voice platforms. You have voice platforms like Siri and Bixby, where you use voice to do the query, but then you get a visual response. And then you can “tap, tap, tap: book” or “tap, tap: buy.” That’s cool, but it’s also great in those moments when tapping isn’t an option – when you’re driving your car or using an Echo. You’re able to simply find — and get — that specific thing you want. And because it’s not a “tap-centric” environment, being able to conduct that transaction via voice is powerful.

Powered by WPeMatico

Speaking To Search Engines: How Marketers Can Drive Better Results In The Intelligent Future

With search engines shifting from delivering blue links on a page to structured answers and knowledge cards, a quickly growing number of searches result in a user never visiting a web page at all — and this means major changes for marketers who may have dedicated the majority of their time to writing website copy.

If the most important audience isn’t a group of people but search engines themselves, how can marketers make sure that their business is providing the right information when customers need it most? At a panel moderated by GeoMarketing‘s Lauryn Chamberlain at Yext’s ONWARD conference, SEO experts discussed this topic — offering tips to marketers on “how to speak search engine” in the intelligent future.

Below, an edited and condensed version of the panel discussion.

GeoMarketing: Let’s start by talking about the knowledge graph. How can marketers work to make sure their business is represented there — and how is the knowledge graph different from featured snippets?

Adam Edwards, head of SEO, US, Reprise: When we talk about featured snippets or featured answers, the best content wins. This really has little to do with structured data; it is controlled by Google programmatically, and it [simply] pulls for the content that answers the search query. If you are structuring your content in a way that conversationally makes sense [on your site], Google is going to serve that to the user as the best answer.

When we talk about the knowledge graph, that is completely structured data driven.  So we [need] to mark that out: We can mark up our logos, we can mark the hours, we can mark up the site.

And Google has evolved this considerably over just the last year: Now we can actually go in and respond to customer questions directly from the knowledge graph. That’s fantastic — especially if you’re a local business — to be able to have that evolution in the knowledge panel, which is a way for you to push offers out and to respond in real time to actual questions from your customers.

Casey Markee, founder, Media Wyse: [It’s important to note] that feedback is very important on featured answers; Google does make mistakes with featured snippets all the time. For example, about three months ago, if you typed in “who owned the New York Jets?” the featured answer was “Tom Brady.”

That was just one of the funniest examples. But in the event that they do get them wrong, definitely avail yourself of the feedback option.

How do you manage your search marketing such that you’re reaching people who already know what they want — e.g. searching “new Camry” and seeing the knowledge card with prices and configurations — versus those who are making a more general search? Are there “search commandments” today, or does it vary?

Casey Markee: When we talk about search commandments, when we talk about qualifying for position zero, when we talk about marketing to the machine… we have to [talk about] JSON-LD schema. It’s a beautiful language, and it’s very easy to learn. Google loves it, and they’ve updated basically every bit of documentation they have online around structured data to say that JSON-LD is a recommended standard so you should be using that.

Google is never going to come out and say, “Hey, you should do this.” But when they say, “This is our recommended standard,” you tend to want to use that.

Also in regards to the knowledge graph, if you go in and look at the help pages from Google, they’ll tell you how to mark that adequately. Just type “knowledge graph” into Google and look at help pages; I’m always shocked at how many SEOs have never visited those pages. Google tells you what they want you to do, and it’s just a matter of you implementing those examples and filling out that material as much as possible.

Brendan King, CEO, Vendasta: To me, it’s really important that everything is marked up so that Google understands it [and so it addresses] what consumers are looking for.

[But] to me, the real challenge is to get [business owners] to understand that they need to get that structured data there, and then you can use all kinds of of tools to make sure it’s marked up correctly — and then after that you still have to find a way to close the loop. What good is the data if it doesn’t actually result in a conversion where [that business] gets some money?

Casey Markee: Right. And I don’t know if I’d use the word “commandment,” but one of the biggest things that I try to instill into my clients is recognizing that we need to get everyone on the same page. Not everyone needs to know everything about the different schema, but they need to know that this is important.

We can’t talk about the future of intelligent search without touching on the role of voice: 20 percent of searches in the Google app are now by voice, and humans are intuitively built to talk and listen. How is voice transforming search, and, as an extension of that, what role do intelligent assistants like Amazon Alexa now play in how brands need to interact?

Casey Markee: You could almost look at voice search as a no user interface — it’s literally like a blank slate. What I mean is, you don’t even see the screen, and so as marketers, we have to predict what people are going to ask with literally very little feedback and you have to invest in that.

It’s not just about having mark up; it’s about having content that will actually answer these questions that users are asking. I’m sure in the future there will be some artificial intelligence solutions, but right now, answering someone’s question is going to come down to just understanding what people could ask.

Adam Edwards: This is also kind of why schema continues to be important: There is going to be a new schema introduced next year that is currently pending right now. It’s called speakable, and it’s specifically optimized around marking up conversational content. So, the idea is that it will allow everyone  to actually pull out snippets and just mark those up with structured data — so that it’s easier for voice assistants to pull that information programmatically from the page in as clean and easy a way as possible.

Brendan King: Yeah, they call it voice “assistance” because it isn’t really just search anymore. It’s interesting the uses that these devices see, and it [brings about] a whole change in your behavior. At my house, the people that use it most are my grandkids — and they’re turning the lights off, playing music, et cetera. [There are ] just so many use cases here.

It really is the future, and the only way that it works is with that structured data.

Adam Edwards: Yeah, I mean, I think it’s now table stakes to have that data out there.

Brendan King: If people hear or see an ad, they expect that they can go immediately look for that business. If they don’t find that business right away, they’re going to find a competitor.

So I like to say the best place to hide a dead body is on the second page of the Google search results, because nobody goes there.

Powered by WPeMatico

Inspiring the American Dream With Intelligent Design

Subscribe For The Latest Updates

Sign up for the Mortgage Geek newsletter and get access to articles curated just for you!